To the Editor: Duloxetine is effective in major depressive disorder (MDD) therapy, but unproven in panic disorder (PD) therapy. Here, we presented five cases of first-onset drug-naive MDD with PD that responded to duloxetine. Their volumes of putamen and brainstem increased after 6 weeks of therapy.

Five patients (3 women, 2 men; mean age: 26.6 years [standard deviation {SD}: 4.56]) with first-onset, drug-naive MDD with PD and five healthy-control subjects (2 women, 3 men; 26.6 [3.36] years old) were enrolled; all signed informed consents, which were approved by the Institutional and Review Board at Buddhist Tzu Chi Hospital, Taipei Branch. They all received clinical rating scales, with the following scores: Hamilton Rating Scale for Depression (Ham-D): 28.8 (4.6) and Panic Disorder Severity Scale (PDSS): 19.8 (4.32). No specific physical illness, psychotic features, past manic episodes, nonpsychiatric medication use, or substance abuse were mentioned. Treatment with duloxetine 30 mg/day was started and titrated to 60 mg within 2 weeks without any significant side effects except mild nausea and sedation. After a 6-week therapy, their MDD and PD symptoms responded to duloxetine (Ham-D: 8.2 [3.83]; PDSS: 3.4 [1.67]).

Structural brain MRI scans were obtained with a 3T GE version scanner housed at Buddhist Tzu Chi Hospital, Taipei Branch. Scans with three-dimensional fast spoiled gradient-echo recovery (3D-FSPGR) T1W1 (TR=11.2 msec, TE=5.2 msec, matrix=256 x 256, field of view=260 mm, number of excitations=1, slice thickness=1 mm, 180 slices; no gap) were performed at the first visit and 6^th-week visit for patients. The controls were also scanned twice, at baseline and 6th week. Structural MRI was processed with FMRIB's Integrated Registration and Segmentation Tool function (FIRST Version 1.2) of FSL (FMRIB Software Library, Version 4.1.1) to perform subcortical brain segmentation using the shape and appearance model.¹ The segmentation files were converted to volume estimation of these structures. We found significant increase of volumes in the putamen and brainstem of these patients (Table 1). There were no significant subcortical changes in healthy controls within 6 weeks, which would exclude the possible bias of artifacts. The changes of putaminal volumes were well correlated with the improvements in Ham-D scores and PDSS scores (Spearman's rho test: r=0.975, two-tailed p=0.005; Spearman's rho test: r=0.949, two-tailed p=0.014). The changes of volumes of brainstem were also correlated with improvements in MDD and PD symptoms (Spearman's rho test: r=0.975, two-tailed p=0.005). The scatter plots showed cluster spots without obvious outliers.

The recent meta-analysis of MR images on MDD patients showed moderate reductions of volumes in the putamen and thalamus.² Sheline has proposed a theory of limbic-cortical-striatal-pallidal-thalamic circuitry in MDD pathogenesis.³ Yoo et al.⁴ reported that PD patients had gray-matter volume decrease and the reduction severity might be correlated with PD symptom severity.⁴ Gorman et al.⁵ proposed a neuroanatomical “fear network” in PD, with projections to the brainstem. The serotonin and norepinephrine dual reuptake-inhibition of duloxetine might restore this potential abnormality.

We thank Dr. Yuan-Yu Hsu, Dept. of Medical Imaging, Buddhist Tzu-Chi General Hospital Taipei Branch, for MRI acquisition help and technical assistance.